Influence of hydrophobization of fumed oxides on interactions with polar and nonpolar adsorbates

Accepted Manuscript Title: Influence of hydrophobization of fumed oxides on interactions with polar and nonpolar adsorbates Authors: M. Gun’ko, E.M. Pakhlov, O.V. Goncharuk, L.S. Andriyko, A.I. Marynin, A.I. Ukrainets, B. Charmas, J. Skubiszewska-Zi˛eba, J.P. Blitz PII: DOI: Reference:

S0169-4332(17)31862-7 http://dx.doi.org/doi:10.1016/j.apsusc.2017.06.207 APSUSC 36407

To appear in:

APSUSC

Received date: Revised date: Accepted date:

22-2-2017 24-3-2017 20-6-2017

Please cite this article as: M.Gun’ko, E.M.Pakhlov, O.V.Goncharuk, L.S.Andriyko, A.I.Marynin, A.I.Ukrainets, B.Charmas, J.Skubiszewska-Zi˛eba, J.P.Blitz, Influence of hydrophobization of fumed oxides on interactions with polar and nonpolar adsorbates, Applied Surface Sciencehttp://dx.doi.org/10.1016/j.apsusc.2017.06.207 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

Influence of hydrophobization of fumed oxides on interactions with polar and nonpolar adsorbates V.M. Gun’ko,a,* E.M. Pakhlov,a O.V. Goncharuk,a L.S. Andriyko,a A.I. Marynin,b A.I. Ukrainets,b B. Charmas,c J. Skubiszewska-Zięba,c J.P. Blitzd a

Chuiko Institute of Surface Chemistry, 17 General Naumov Street, 03164 Kyiv, Ukraine

b

National University of Food Technology, 68 Volodymyrska Street, 01033 Kyiv, Ukraine

c

Faculty of Chemistry, Maria Curie-Skłodowska University, 20-031 Lublin, Poland

d

Eastern Illinois University, Department of Chemistry, Charleston, IL 61920 USA

∗Corresponding

author. Tel.: +38044 4229627; fax: +38044 4243567.

E-mail address: [email protected] (V.M. Gun’ko).

Graphical abstract

Highlights 

Textural characteristics of modified fumed oxides are affected by modification degree and modifier type



Interfacial behavior of water depends more strongly on surface modification than that of n-decane



Confined space effects appear for both polar (water) and nonpolar (decane) adsorbates

Abstract A variety of unmodified and modified fumed silica A-300 and silica/titania (ST20 and ST76 at 20 and 76 wt.% of titania, respectively) was prepared to analyze features of their interactions with polar and nonpolar adsorbates. The materials were studied using nitrogen adsorption-

2 desorption, ethanol evaporation kinetics, infrared (IR) spectroscopy, thermogravimetry (TG), photon correlation spectroscopy, differential scanning calorimetry (DSC), DSC and TG thermoporometry, and quantum chemistry. Changes in surface structure of modified nanooxides with increasing hydrophobization degree (MS) from 20% to 100% have a strong affect on the textural characteristics of the materials and adsorption-desorption of various adsorbates. Confined space effects enhanced due to the location of adsorbates in narrow voids between nanoparticles lead to freezing-melting point depression for bound polar and nonpolar adsorbates. The behavior of particles of modified nanooxides in aqueous and water/ethanol media is strongly altered due to enhanced aggregations with increasing value of MS. All of these change are non-monotonic functions of MS which affects (i) rearrangement of nanoparticles, (ii) interactions with polar and nonpolar adsorbates, (iii) location of adsorbates in voids of different sizes, (iv) the clustering of adsorbates and formation of nearly bulk structures.

Keywords: Hydrophilic nanosilica; Hydrophobic nanosilica; Textural characteristics; Interfacial phenomena 1.

Introduction Both totally and partially hydrophobized fumed metal or metalloid oxides (FMO) such as silica,

titania, silica/titania, etc. are of practical interest. These materials can used as improved lyophilic fillers in nonpolar polymers and other applications, being more versatile than the initial hydrophilic FMO. These uses have been previously described in detail in monographs [1-10] and by FMO producers [11-14]. There are various low-molecular weight modifiers which can be used to hydrophobize FMO such as chlorosilanes (ClxSiR4-x, where R = CH3 or other organic functional groups), organosilanes ((CH3O)xSiR4-x), hexamethyldisilazane, etc.). Another pathway to hydrophobize the surface of FMOs is to use poly(dimethylsiloxane) (PDMS), in the presence of dimethyl carbonate (DMC) [15-17]. This modification provides a much thicker functional layer than simple silanes produce. The physicochemical properties and characteristics of modified silicas depend on the amount, surface distribution and structure of attached functionalities, the degree of substitution of surface OH groups by di- or trimethylsilyl groups (MS), silanes with longer functionalities, and fragments of depolymerized PDMS, etc. [1-17]. Features of the initial silica including particle size distribution, specific surface area (SBET), and porosity, can have profound affects on the properties of the final materials [1-8]. It is also important that silane-modified FMOs remain highly disperse materials with only slightly decreased SBET values compared to the original unmodified materials.

3 Functionalized FMO can be an important component in various composites [18-26] including superhydrophobic surfaces [25-27]. Surface functionalization affects interactions of silicas as well as silica filled polymers with water [28-35] and other solvents [1-12,36,37]. Surface functionalization can also affect chromatographic [1,10,38] and catalytic [9,11,39] efficiency. The properties of silicas and related composites depend on the organization of secondary (aggregates of primary nonporous nanoparticles, NPNP) and tertiary (agglomerates of aggregates) structures [40-42]. Features of these structures depend on the morphology and chemical composition of NPNP which dictate the types of inter-particle interactions at play. When present, these interactions are also influenced by the composition of a functional layer at a NPNP surface. The uniformity or non-uniformity of the distribution of surface functionalities on partially modified FMO influence the properties of the final composites with FMO filled polymers [1-39]. Therefore, it is of interest to compare the effects of hydrophobization of nanooxides with dimethylsilyl groups with crosslinks, non-cross linked trimethylsilyl groups, and poly(dimethylsiloxane) (PDMS) depolymerized with dimethyl carbonate (DMC) on the interactions with polar and nonpolar adsorbates. There are no papers in the literature devoted to similar comparisons of a variety of unmodified and modified nanosilicas and titania/silicas interacting with polar and nonpolar adsorbates. Thus, the main focus of this work is on the interfacial phenomena at surfaces of differently modified FMO vs. the content and type of hydrophobic surface functionalities.

2.

Materials and methods

2.1. Materials The first series of nanosilicas (M0-M5, Table 1, Pilot plant of Chuiko Institute of Surface Chemistry (PPCISC), Kalush, Ukraine) with various degrees of silanol substitution by crosslinkable dimethylsilyl groups was prepared using the reaction of nanosilica A300 (labeled as M0) with various amounts of hydrolyzed dimethyldichlorosilane (DMS) using certain amounts of water. The second series of silicas (CM1-CM4, Table 1) was prepared with A300 (PPCISC) as the initial material, subsequently modified using different amounts of non-cross-linkable hexamethyldisilazane (HMDS) [40,41]. The silica modification was carried out with HMDS dissolved in n-hexane in a glass reactor (volume 1 dm3) with a mixer (500 rpm) at 100 oC. After the reaction, the samples were treated in air at 95 oC for 1 h to remove n-hexane and residual NH3.These functional groups were not cross-linked. The third series of silicas (B1-B3, AM1, Table 1) was prepared as a mechanical blend of unmodified A-300 with totally hydrophobized AM1 (DMS modified A300 from PPCISC) at

4 weight ratios 3 : 1, 1 : 1, and 1 : 3; respectively. These samples were prepared by reactor treatment in an ethanol slurry (2 ml per 10 g of silica) at room temperature for 2 h. This series provides a comparison of materials with a random but more uniform distribution of hydrophobic functionalities (primarily the first “M” series, and also the second “C” series of samples with all NPNP partially covered by silanes), with samples exhibiting a decidedly non-uniform distribution of hydrophobic functionalities (the third “B” series consisting of a mixture NPNP completely covered and completely uncovered with silanes). Thus, the third “B” series of materials is used to model samples with maximal non-uniformity of modified silica since some nanoparticles are pure unmodified silica, while others are completely modified. The fourth series includes silica/titania at CTiO2 = 20 wt.% (ST20) and 76 wt.% (ST76) (PPCISC) modified to completely replace surface OH groups by hydrophobic functionalities using either HMDS (“Siliconpolymer”, Zaporozhye, Ukraine), or poly(dimethylsiloxane) (PMS-200, molecular weight 70001000) in the presence of dimethyl carbonate (PDMS/DMC). The FMO surface modification using PDMS/DMC leads to depolymerization of PDMS, breaking of surface SiO and TiO bonds, and attachment of the PDMS fragments to the surface. Reaction of HMDS (1.1 ml/2 ml of n-hexane for ST20 or 2.2 ml/4 ml of n-hexane for ST76) with ST (10 g of ST20 or 20 g of ST76) was carried out in a glass reactor (0.5 dm3) with a mixer. The reaction mixture was stirred at 20 oC for 1 h, then temperature was increased to 100 oC and the mixture was stirred for an additional 1 h. After reaction, the samples were dried in an oven at 90 oC in air to remove hexane and residual NH3. A mixture of PDMS with DMC (1 : 1 v/v) was added (8 ml, dropwise) to both ST (20 g) in the glass reactor. The slurry was mixed for 1 h at room temperature, then the temperature was increased to 200 oC over 2 h and maintained at this temperature for an additional 1 h. The samples were subsequently cooled to room temperature. Fumed alumina/silica/titania, AST (PPCISC) with component amounts in wt. % of 21 (alumina), 8 (silica), and 71 (titania) (AST71) was used in comparative DSC measurements. The initial fumed oxides as well as some of these surface functionalized materials and their properties have been described in detail elsewhere [5-7,29,42-45]. These studies and others have shown that the degree of hydrophobicity of modified silicas plays an important role in material properties for various applications [1-28,46-48].

2.2. Textural characteristics To analyze the textural characteristics of nanooxides degassed at 353 K or 373 K for several hours (Table 1), low-temperature (77.4 K) nitrogen adsorption–desorption isotherms (Fig. S1, Electronic Supplementary Information (ESI) file) were recorded using a Micromeritics ASAP 2420

5 adsorption analyzer. The specific surface area (SBET) was calculated according to the standard BET method [49]. The total pore volume Vp was evaluated from the nitrogen adsorption at p/p0  0.980.99, where p and p0 denote the equilibrium and saturation pressure of nitrogen at 77.4 K, respectively [50]. It should be noted that the value of Vp is much lower than the empty volume (Vem = 1/b1/0, where b and 0 are the bulk and true densities of samples) in the FMO powder, since Vem can reach 25 cm3/g for A-300 at b  0.04 g/cm3, but the value of Vp is typically less than 1 cm3/g (see Table 1). This underestimation of the value of Vp calculated from the nitrogen adsorption is due to a very weak influence of the pore walls (nanoparticle surface) on nitrogen molecules located in macropores far from the solid surface, in loose FMO agglomerates. The nitrogen desorption data were used to compute the pore size distributions (PSD, differential fV(R) ~ dVp/dR and fS(R) ~ dS/dR) using a self-consistent regularization (SCR) procedure under non-negativity condition (fV(R)  0 at any pore radius R) at a fixed regularization parameter  = 0.01. A complex pore model was applied with slit-shaped (S) and cylindrical (C) pores and voids (V) between spherical NPNP packed in random aggregates (SCV/SCR method) [51]. The differential PSD with respect to the pore volume fV(R) ~ dV/dR, fV(R)dR ~ Vp were re-calculated to incremental PSD (IPSD) at V(Ri) = (fV(Ri+1) + fV(Ri))(Ri+1  Ri)/2 at V(Ri) = Vp). The fV(R) and fS(R) functions were also used to calculate contributions of nanopores (Vnano and Snano at 0.35 nm < R < 1 nm), mesopores (Vmeso and Smeso at 1 nm < R < 25 nm), and macropores (Vmacro and Smacro at 25 nm < R < 100 nm) [51]. The values of and as the average pore radii were calculated as a ratio of the first moment of fV(R) or fS(R) to the zero moment (integration over the 0.35-100 nm range) = f(R)RdR/f(R)dR.

2.3.

(1)

HRTEM and SEM High resolution transmission electron microscopy, HRTEM (JEM–2100F, Japan) images

were recorded for initial A-300 as a representative sample. A powder sample was added to acetone (chromatographic grade) and sonicated. Then a drop of the suspension was deposited onto a copper grid with a thin carbon film. After acetone evaporation, sample particles that remained on the film were studied with HRTEM. The morphology was analyzed using scanning electron microscopy (SEM) with a DualBeam Quanta 3D FEG FEI apparatus under conditions of low vacuum at an accelerating voltage of 2-10 kV.

2.4. Infrared spectroscopy

6 The infrared (IR) spectra over the 4000–300 cm1 range (at 4 cm1 resolution) were recorded in transmittance mode using a Specord M80 (Carl Zeiss, Jena) spectrometer using sample powders mixed with KBr (1 : 400), treated in a grinder (with a stainless steel sphere of 10 cm3 in volume and a stainless steel ball of 0.8 cm in diameter, 30 W, and frequency 50 Hz) for 5 min to prepare a uniform blend, and then pressed at 9×103 atm. to form thin tablets (~20 mg) that provides a transmittance of > 10 % in the range of maximal absorption from SiOSi asymmetric stretching vibrations at 1100 cm1 for silica [4,52]. The transmittance spectra were re-calculated into absorbance spectra for quantitative analysis. IR spectra used to analyze the OH and CH stretching vibrations were recorded using thin pressed pellets without KBr.

2.5. Ethanol evaporation For the first three series of samples (Table 1), evaporation of ethanol (0.4965990.055825 g per 0.0502330.000837 g of silica) was studied at 181 oC using an ABT 220-5 DM (Kern, Germany) analytical balance. Mass loss from ethanol evaporation was monitored for ca. 50 h. To reduce errors caused by changes in environmental conditions, the evaporation of ethanol from open glass vials (volume 10 cm3) was carried out simultaneously for all samples. Ethanol was selected as a probe compound because it interacts well with both hydrophilic and hydrophobic silicas [48].

2.6.Particle sizing Particle size distributions of silicas were studied using a Zetasizer Nano ZS (Malvern Instruments) apparatus. Distilled water or water/ethanol mixtures (1:1) with 1 wt.% silica were utilized to prepare suspensions that were sonicated for 2 min using an ultrasonic disperser (BANDELIN electronic НВ 2070, power 70 W and frequency 20 kHz) prior to analysis.

2.7.

DSC

Differential scanning calorimetry (DSC) investigations of interactions of unmodified fumed oxides with nonpolar (n-decane) and polar (water) adsorbates were carried out using a PYRIS Diamond (Perkin Elmer Instruments, USA) differential scanning calorimeter calibrated at different heating rates using Perkin Elmer recommended calibration procedures. DSC is sensitive to phase transitions, and is thus used in thermoporometry to study the textural (pore size distributions) characteristics of a variety of fumed oxides (see ESI file).

2.8. Quantum chemical calculations

7 Quantum chemical (QC) calculations using the density functional theory (DFT) method were carried out using a hybrid functional ωB97X-D [53-55] (labelled as wB97XD in Gaussian 09) with the cc-pVDZ basis set with the Gaussian 09 program suite [53]. Solvation effects were analyzed using the solvation method SMD [56] implemented in Gaussian 09. To compute the Gibbs free energy of solvation (subscript s), ΔGs = Gl  Gg, where Gl and Gg are the Gibbs free energies of a molecule free or bound to a silica cluster in the liquid (subscript l) and gas (g) media, respectively. Calculations were performed taking into account zero-point and thermal corrections to the Gibbs free energy in the gas phase and for molecules and silica clusters using the geometry optimized with HF/cc-pVDZ or ωB97X-D/cc-pVDZ [53] or 6-31G(d,p) and calculated taking into account the solvation effects (using SMD). The latter basis set was used only for the initial silica cluster [57]. Solvation effects were analyzed for the silica clusters with 8 tetrahedra comprising various numbers of OH and TMS groups (see ESI file). Note that the functional ωB97X-D introduces empirical damped atom-pairwise dispersion terms containing range-separated Hartree-Fock exchange for a better description of van-der-Waals (vdW) interactions [53-58]. Therefore, this functional was selected to obtain improved results for the gas and liquid (SMD) phases with both hydrophilic and hydrophobic solvents.

3. Results and discussion In general the focus here is to analyze the characteristics and properties of differently modified FMO and their influence on the interactions with polar and nonpolar adsorbates. Therefore, some methods used (e.g., QELS and ethanol evaporation) may appear inapplicable for the study of these materials, but they allow one to elucidate some aspects of the interactions of modified FMO with selected polar and nonpolar adsorbates. 3.1. Textural characteristics vs. silylation and surface structure In the IPSD (Fig. 1) calculated using the nitrogen adsorption-desorption isotherms (see Fig. S1 in ESI file) treated with the SCV/SCR method, the first peak at R = 1-2 nm corresponds to voids between adjacent nanoparticles in the same aggregate. The second PSD peak at R = 10-30 nm is due to voids between nanoparticles without direct contacts in the same aggregate, or neighboring aggregates in agglomerates [42]. The third PSD peak or a shoulder at R = 30-90 nm can be due to voids in aggregates and between aggregates in agglomerates. Thus the textural pores, which result from voids between nanoparticles in secondary structures in the nanooxide powder, demonstrate a broad IPSD. This is important in explaining effects on the interfacial phenomena studied; specifically that the IPSD exhibits a weak dependence on surface hydrophobicity when modified by relatively small DMS or TMS groups (Fig 1a-c). In

8 contrast, use of the larger PDMS/DMC modification with fumed silica/titania has a more profound effect on IPSD resulting in more complex changes (Fig. 1d). The weight of attached PDMS/DMC is several times larger than both DMS and TMS (see TG data in ESI file). Increasing the silica surface content of dimethylsilyl (DMS) groups (series M0-M5, Table 1, MS, Fig. 1a), trimethylsilyl (TMS) groups (series CM1-CM4, Fig. 1b) and AM1 in the blends with A-300 (series B1-B3, Fig. 1c) all result in decreasing specific surface area (SBET). Also noted is a decrease in the contribution of nanopores at R < 1 nm (Snano, Vnano) because of blocking of narrow voids between adjacent NPNP by surface functionalities. This also results in an increase in effective size of the modified nanoparticles. For certain samples of the first and third series with cross-linked functionalities, i.e., with DMS groups, the pore volume (Table 1, Vp) decreases with increasing MS due to both increasing nanoparticle size and secondary structure reorganization after the silica modification reaction. Similar textural changes are observed resulting from modification of ST20 and ST76 (Table 1, Fig. 1d). The values of S and V of mesopores at 1 nm < R < 25 nm and macropores at R > 25 nm, as well as the average values of the pore radii with respect to the pore volume and surface area , are non-monotonic functions of MS (Table 1). Relatively complex and often contradictory forces can result in the IPSD functions seen at R > 1 nm (Fig. 1). Any modification of nanosilica leads to (i) changes in the geometry of nanoparticles; and (ii) significant changes in the arrangement of nanoparticles in their aggregates and agglomerates of aggregates which are relatively ‘soft’ [42,45]. These changes can be opposing for pores in different ranges of pore size. For example the IPSD peaks at 1-2 nm and 10-30 nm shift in opposite directions with increasing MS for the first and third series of silicas (Fig. 1a,c), with different results for the second series (Fig. 1b). For complex FMO such as silica/titania, there are several additional factors which affect both the surface modification results and subsequent changes in material properties. For example, residual water remaining in ST20/HMDS and ST76/HMDS is located in narrow pores (nanopores and narrow mesopores at R < 4 nm according to thermoporometry calculations, see ESI file). The pores with water are narrower than pores filled by nitrogen molecules. This is due not only to the smaller size of H2O than N2, but also due to penetration of water molecules into fumed oxide nanoparticles. This adsorbed water can hydrolyze TiOSi bonds upon attachment of silane functionalities to the ST surface. Nanosilicas swell upon interaction with water [1,2,4,42]. However surface hydrophobization can result in significantly reduced swelling. Unlike nanosilicas, ST materials contain nanocrystallites of titania which do not undergo swelling, as well as amorphous titania and silica phases which do swell. The more complex ST materials (compared to nanosilicas) leads to unique

9 features in surface hydrophobization results, especially at the titania phase surface and the silicatitania boundary, resulting in changes in the textural characteristics of modified FMO. Surface modification of ST samples by PDMS/DMC resulting in surface bond breaking on NPNP results in these materials exhibiting maximal deviations from the pore shape model (Table 1, w), as well as the greatest losses in SBET surface area values. Surface modification by DMS (Fig 1a, Table 1 M samples) or TMS (Fig. 1b, Table 1 CM samples) groups lead to changes in textural characteristics, however they are not very significant in contrast to that of ST modified by PDMS/DMC. Structural and textural features of unmodified and modified FMO lead to some characteristic changes in IR spectra (Figs. S2-S5), desorption processes studied using thermogravimetry (TG, Figs. S6-S10), and evaporation of ethanol (Fig. S11), due to changes in the polarity of surface functionalities (Fig. S12). This also leads to a significant difference in the temperature behavior of polar (water) and nonpolar (n-decane) adsorbates bound to unmodified and modified FMO (vide infra). Changes in surface structure are clearly visible in the IR spectra of samples as a function of degree of surface modification (Figs. S2-S5). An increase in the amount of DMS or TMS groups leads to a decreased intensity of the O-H stretching vibrations of free silanols [4,52] characterized by a narrow band at 3750 cm1 as well as adsorbed water (a broad band at 3700-2700 cm1), but increased intensity of C-H stretching vibrations (3000-2900 cm1). Note that intensity of a silanol band at 3660 cm1 changes much less than the 3750 cm1 free silanol band. The former is attributed to silanols that are less accessible [4] to silane molecules during the surface modification. For samples with the largest content of hydrophobic functionalities (M5, CM4, AM1), the free silanol band at 3750 cm1 is no longer detected. However, this does not prevent water adsorption since the intensity of a broad band at 3600-2700 cm1 is detectable for these samples (Fig. S5). There are small localized regions between DMS or TMS groups around siloxane bonds that exhibit a negative potential (Fig. S12 in ESI file), where small water clusters can adsorb. However, despite this adsorbed clustered water, the samples at MS > 50% cannot be wetted in aqueous media. The approach previously developed for silicas [59] applying the formula SIR,OH = 299.48874*IOH/ISiOSi  18.29504 (where IOH and ISiOSi are integral intensity of the bands at 3760-3710

(2) cm1

and 1940-1760 cm1,

respectively) was used to estimate the total specific surface area of unmodified silicas, and the contribution of the unmodified portion of a surface of partially modified silicas (see Table S1, SIR,OH in ESI file). Additionally, the shape of a complex band between 1200-1000 cm-1 corresponding to longitudinal (LO), transverse (TO), and surface optical (SO) phonon modes of

10 the SiOSi asymmetrical stretching vibrations depends on the diameter (d) of silica nanoparticles. A correlation function [60] linking the intensity of surface modes at 1200 cm1 and the value of d was used to estimate the specific surface area (Table S1, SIR,SiOSi) of both unmodified and modified silicas. Since the intensity of surface modes in the SiOSi stretching vibrations in DMS and TMS modified materials is much greater than that of the surface SiOSi bonds of silica nanoparticles, overestimation of the value of SIR,SiOSi occurs. Also noted is that the value of SIR,OH is not zero for the M5, CM4, and AM1 samples (Table S1), perhaps due to the presence of a small amount of residual groups, or an artifact of integration in the 3760-3710 cm1 range for these completely modified silicas (Figs. S2-S5). The thermal stability of a surface functional layer on modified FMO surely depends on both the types of surface groups and the oxide. According to thermogravimetry data (TG/DTG/DTA) (Figs. S6-S9 in ESI file) and TG thermoporometry (Fig. S10), thermal decomposition of the functional groups starts at a lower temperature for CM4 with TMS groups than M5 with DMS groups, or PDMS/DMC at the ST surface which has residual adsorbed water despite hydrophobization. An increase in titania content in ST76 compared with ST20 leads to a decrease in temperature of maximum decomposition of PDMS functional groups (429 and 468 oC, respectively). Note that maximum water desorption is observed at a lower temperature for ST76/HMDS than ST20/HMDS (Fig. S8, at 97 and 101 oC, respectively). The desorption processes in air are continue to occur for all samples up to 1000 oC due to several reactions such as desorption of intact water, associative desorption of water due to condensation of surface hydroxyls, as well as decomposition and oxidation reactions of the methylsilyl groups. Associative desorption of water (i.e. condensation of residual hydroxyls) from the surface of nanoparticles, or removal of water (both intact molecules and condensed hydroxyls) from the volume of these particles [1-7,42] can occur at temperatures higher than MS decomposition temperatures. Thus, residual water and hydroxyls remain even in completely hydrophobic samples. These samples contain few if any free hydroxyls demonstrated by the lack of an IR band at 3750 cm1, but the presence of water and disturbed hydroxyls as a broad IR band at 3700-2700 cm1 is detected (Fig. S5). Note that the TiOSi(CH3)3 bonds are hydrolytically unstable, resulting in removal of TMS groups from the TiO2 patches at the ST surface [42]. There is more residual water on ST/HMDS than on ST/PDMS/DMC resulting from enhanced screening of

TiOSiR groups by the long

functionalities of PDMS fragments inhibiting water adsorption and reaction with TiOSiR groups.

3.2.

Interfacial phenomena

Commented [JB1]: Not clear to me what MS stands for.

11 The surface modification of nanosilicas changing the nature of the surface (Fig. S12, Table S2 in ESI) alters the behavior of these materials in liquid media which is of practical importance [1-14]. When the degree of surface modification is greater than 50% it is impossible to wet these silicas by liquid water. Therefore, all the samples were studied in water/ethanol (1:1) media, and when possible partially modified samples were studied in aqueous media (Fig. 2). Aggregation of both unmodified and partially modified samples in aqueous media is less than in the water/ethanol mixture (Fig. 2). Surprisingly unmodified silicas exhibit the greatest difference in aggregation behavior, measured by particle volume, between the two liquid media. There are no individual (non-aggregated) nanoparticles detected in any of the suspensions. Aggregates of nanoparticles (< 1 m) [42] are not observed at high MS values, only agglomerates (> 1 m) of aggregates are observed. Interestingly, M4 (not M5) exhibits the largest agglomerate size for this first series of samples. For the second (Fig. 2c,d) and third (Fig. 2e,f) series, the largest size of agglomerates is observed for samples CM4 and AM1 corresponding to samples with the greatest MS values. The most non-uniform particle size distributions (PaSD) are observed for samples with the smallest amount of hydrophobic functionalities (Fig. 2, M1 and CM1), or in a mixture of AM1/A-300 with a minimal amount of AM1 (Fig. 2e, B1) in both liquid media. These NPNP surfaces with low hydrophobic surface content exhibit the most surface non-uniformity [42], resulting in non-uniform particle size distributions. Note that the PaSD with respect to particle number appear more uniform than those related to particle volume because for the former, the number of small particles have a greater effect on the PaSD because of the dependence of the scattering effects on particle sizes. Evaporation kinetics was used to analyze the interaction of ethanol with unmodified and modified silicas. The interfacial area and structure of a surface layer affect the evaporation kinetics of ethanol from concentrated suspensions (0.05 g of silica per 0.5 g of ethanol). Figure 3 shows the final portion of these evaporation kinetics curves, S11 shows the complete curves. It is difficult to establish features of the evaporation kinetics (Fig. S11) without taking into account differences in the specific surface area of the silicas (Table 1, SBET). The picture becomes clear after normalizing the evaporation kinetics data by dividing mev/m0 by SBET vs. time (Fig. 3). The greater the content of hydrophobic functionalities, the faster the evaporation for all series of samples. However, the changes in the evaporation rate are nonmonotonic because of changes in the organization of secondary and ternary silica particles with increasing amounts of hydrophobic functionalities (Figs. 1, 2, S5, Table 1).

12 The solvation energy of functionalized silicas in both ethanol and water decreases (by the modulus) with increasing degree of surface functionalization (Fig. 4). In contrast in nonpolar hexane the solvation energy increases slightly with increased surface functionalization. The evaporation rate of ethanol increases with increasing silica hydrophobicity (Fig. 3) due to reduction of the interaction energy of ethanol with modified silicas, consistent with Fig. 4. For individual particles, there is a linear decrease vs. TMS (Fig. 4). However, this effect becomes more complex for real samples due to changes in the organization of secondary and ternary particles vs. MS (Fig. 2), which also depends on the type of dispersion medium.

It should be noted that in pores with a hydrophobic surface, air bubbles could remain upon wetting or suspending in aqueous media [42]. However, this effect is negligible modified FMO wetted in ethanol. Therefore, ethanol was used to study evaporation from modified FMO (Fig. 3). For deeper insight into the interfacial phenomena of unmodified and modified FMO surfaces, the interactions of water (Figs. 5 and 7, Table S3) and n-decane (Figs. 6 and 7, Table S4) with a variety of samples were studied using the DSC method. It can be generally stated that water should be more sensitive to surface structure due to specific interactions with various polar surface functionalities than decane, which is limited to van der Waals interactions. Thus DSC measurements with respect to adsorbate amounts, nanooxides structure, and type of modification exhibit larger changes with water compared to decane. For example, the range of maximum values for enthalpy of freezing is 231.4 J/g for water and 39.3 J/g for decane; and the range of maximum values for the enthalpy of melting is 278.7 J/g for water and for decane is 18.2 J/g (Tables S3 and S4 for original data). Note that the enthalpy of fusion at the freezing point (333.5 (water) and 201.82 (n-decane) J/g) for the individual compounds [61] is not so different. Thus the large differences observed for water, and comparatively modest differences found for decane adsorbates, is due to the much greater differences in solvation effects with increasing hydrophobization for water. This difference is observed even for water and ethanol (Fig. 4). Therefore, the effects of freezing-melting point depression depend on water content and surface structure much more than decane (vide infra, Figs. 5-7). In other words, the amounts of strongly bound water (SBW) decreases with increasing surface hydrophobicity and the relative contribution of weakly bound water (WBW) increases. A thicker interfacial water layer senses the influence of a silica surface more than decane. This is due to the stronger hydrogen bonding interactions of water molecules with the silica surface compared with the dispersion interactions of nonpolar decane molecules with the surface (Fig. S12,

13 Table S2). Therefore, the melting point endotherms of frozen bound water shift toward 0 oC with increasing MS value, i.e. the contribution of WBW increases and SBW amount decreases (Fig. 5b,d,f). In contrast, the endotherm melting point peak position of decane is practically independent of the MS value (Fig. 6b,d,f). The position of this peak does not depend on the FMO structure (silica or TS), as well as on the structure of functional groups at a NPNP surface (Figs. 5-7). Thus decane is a rather weakly bound adsorbate. These features of the interfacial behavior of bound water and decane appear in the PSD calculated using DSC thermoporometry (Fig. 8).

Decane layers located far from the silica surface are not influenced by the surface similar to nitrogen adsorption in macropores. Therefore, decane melting occurs at T  Tm (due to kinetic o

effects upon heating at a rate of 10 C/min), therefore this fraction of decane is ignored on the PSD calculations using the DSC data. Since the amount of decane influenced by the surface is relatively small for all FMO studied, the PSD based on DSC curves of decane are located in the same range; mainly at R = 2.0-2.6 nm. Voids (R < 2 nm, Fig. 1) related to closely located nanoparticles in aggregates are poorly accessible for decane molecules which are 10x larger than nitrogen molecules. Thus the effects of decane freezing and melting in narrow voids at R < 2 nm are not observed.

Some shifts of decane freezing point exotherms for FMO studied (Figs. 6 and 7) are due to differences in the amounts of adsorbate (Table S4, Vads = 0.59-3.54 cm3/g). The lower the adsorbate content, the lower freezing temperature. This is due to stronger confined space effects. The freezing exotherms observed at Tm,decane < T < 0 oC are due to residual amounts of adsorbed water, since samples were prepared in air for DSC measurements. The PSD (Fig. 8) calculated using DSC thermoporometry show only a portion of pores filled by unfrozen adsorbates at T < Tm due to confined space effects resulting in freezing-melting point depression. These PSD allow one to obtain information about changes in location of adsorbates vs. MS rather than information on the real textural characteristics. Note that these PSD depend on the amounts of adsorbates (Tables S3 and S4, Vads) which varied from 0.16 to 4.15 cm3/g and from 0.59 to 3.54 cm3/g for water and decane, respectively. There are several tendencies to consider: (i) the smaller the adsorbate amount, only the narrower the pores filled by this adsorbate unfrozen at T < Tm due to stronger confined space effects in narrower pores; (ii) despite lowering amounts of bound adsorbates with increasing degree of hydrophobization, water tends to form large structures consistent with enhanced aggregation of more strongly modified silicas in liquid media; (iii) reorganization of bound decane is very weak due to weak interactions with initial or modified

Commented [JB2]: Discussed above?

14 FMO nanoparticles; (iv) melting effects for water located in nanopores at R < 1 nm or for decane at R < 2 nm are not observed by DSC due to very low heat effects. However, similar effects for water are observable by low-temperature 1H NMR spectroscopy [42] since this method exhibits higher sensitivity due to the detection of only mobile molecules. For example, the 1H NMR spectrum of water bound to partially silylated nanosilicas (Fig. S13, curve 1) demonstrates the clustered structure of water. In addition to bulk water with a signal at H = 3.5-6 ppm, there are two signals with lower H values characteristic of weakly associated water. Similar features are observed in the theoretical spectrum of water bound to two partially silylated nanoparticles (Fig. S13, curve 2); strongly differing from the spectrum of a nanodrop (~3 nm in size) of free water (Fig. S13, curve 3). Formation of clusters (1-2 nm in size) and domains (10-30 nm) of unfrozen water (Fig. 8), is accompanied by the formation of large domains (> 30 nm) characterized by properties similar to bulk water that melt at T > 0 oC (Figs. 5 and 7). Thus, greater degree of hydrophobization results in lower confined space effects and increased aggregation of modified nanoparticles in liquid media. Thawing of frozen decane is observed at T > Tm,decane due to the kinetic delay of the process. The contribution of a bulk, non-bound fraction of adsorbates (NBA) depends on adsorbate amount, material texture, and surface structure of FMO, e.g. the degree of surface hydrophobization. The ratio of endotherm area (and H values, see ESI file) at T < Tm (bound adsorbate) and T > Tm (nonbound bulk adsorbate) show that the contribution of NBA is low for all samples. In aqueous media, 10-12 wt.% of initial nanosilica A-300 can disturb nearly all of the water according to 1H NMR spectroscopy data [42]. However, DSC measurements indicate that the relative amounts of disturbed adsorbates are much lower. This is due to the lower sensitivity of the DSC method detecting all adsorbates, in contrast to low-temperature 1H NMR spectroscopy which detects only mobile adsorbates.

4. Conclusion Partial and complete hydrophobization of fumed silica or silica/titania leads to detectable, small changes in the textural characteristics of modified silicas when the surface functionalities are relatively small dimethylsilyl (which can be cross-linked) or trimethylsilyl (non-cross-linked) groups. However, surface modification by PDMS depolymerized by dimethyl carbonate with breakage of surface siloxane bonds results in much larger changes in textural characteristics. This is due to a larger size of the PDMS/DMC layer, which is several times greater than DMS or TMS. Besides these changes, organization of secondary and ternary particles, such as aggregates of nanoparticles and agglomerates of aggregates, exhibit changes resulting from the modification

15 reaction. This effect increases with increasing content of surface hydrophobic functionalities because the inter-particle interactions of initial hydrophilic and hydrophobic modified nanoparticles strongly differ in gas and liquid media. The structural changes lead to changes in the behavior of modified nanooxides in aqueous or water/ethanol media, corresponding to enhanced aggregation of particles with increasing surface hydrophobization. The most nonuniform particle size distributions are observed for partially modified silicas with a relatively low content of hydrophobic functionalities. The hydrophobization results in reduced interactions of nanoparticles not only with water, but also with nonpolar n-decane and polar ethanol. Therefore, the evaporation rate of ethanol increases with increasing degree of hydrophobization of silicas. Not only is the degree of surface hydrophobization important, but also the type and length of surface functionalities that allow one to control the properties of the modified FMO. DSC measurements of the interactions of water and decane, resulting in data of size distributions of pores filled by unfrozen (at T < Tm) adsorbates, provide information about the structure of bound liquids as a function of many parameters. These parameters include the degree and uniformity of hydrophobization, the types of FMO adsorbents and surface functionalities, and adsorbate concentration. Despite the lower the content of adsorbate, and the narrower the pores filled by this adsorbate, a fraction of adsorbate can form large domains characterized by the properties close to that of bulk liquids. DSC or TG thermoporometry are methods to analyze the structure of bound liquids rather than the textural characteristics of the hydrophobized materials. Liquids can fill pores incompletely in which nanobubbles of air can remain, especially in adsorbents hydrophobized by long functionalities and polar liquid (e.g. water) or nonpolar liquid but with relatively large molecules (e.g. decane). We are unaware of any similar analysis comparing several series of fumed silica and silica/titania with different degrees of hydrophobization by different compounds. Comparison of several sets of differently modified FMO allows us to establish certain features and trends in the textural characteristics and interactions with polar and nonpolar adsorbates. These trends differ depending on the amounts and types of hydrophobic functionalities, as well as the types of initial FMO matrices. A similar analysis of several sets of different materials modified with different functionalities could be useful for an improved understanding of the interfacial phenomena at functionalized surfaces. This may be of importance since such materials are of use for various applications in which properties can be tailored by not only a controlled degree of hydrophobization, but also controlled changes in textural and other material characteristics.

Acknowledgment

16 The authors are grateful to European Community, Seventh Framework Programme (FP7/2007–2013), Marie Curie International Research Staff Exchange Scheme (IRSES grant No 612484) for financial support.

17

References [1]

R.K. Iler, The Chemistry of Silica, Wiley, Chichester, 1979.

[2]

H.E. Bergna, W.O. Roberts (Eds.), Colloidal Silica: Fundamentals and Applications, CRC Press, Boca Raton, 2006.

[3]

E.F. Vansant, P. Van Der Voort, K.C. Vrancken, Characterization and Chemical Modification of the Silica Surface, Elsevier, Amsterdam, 1995.

[4]

A.P. Legrand (Ed.), The Surface Properties of Silicas, Wiley, New York, 1998.

[5]

A. Dabrowski, V.A. Tertykh (Eds.), Adsorption on New and Modified Inorganic Sorbents, Studies in Surface Science and Catalysis, Vol. 99, Elsevier, Amsterdam, 1996.

[6]

J.P. Blitz, V.M. Gun'ko (Eds.), Surface Chemistry in Biomedical and Environmental Science, NATO Science Series II: Mathematics, Physics and Chemistry, Springer, Vol. 228, Springer, Dordrecht, 2006.

[7]

A.A. Chuiko (Ed.), Chemistry of Silica Surface, UkrINTEI, Kiev, 2001 (in Russian).

[8]

N. Auner, J. Weis (Eds.), Oganosilicon Chemistry VI, Wiley-VCH, Weinheim, 2005.

[9]

Ullmann's Encyclopedia of Industrial Chemistry, Wiley-VCH, Weinheim, 2008.

[10]

K.H. Büchel, H.-H. Moretto, P. Woditsch, Industrial Inorganic Chemistry, Wiley-VCH, Weinheim, 2000.

[11]

Basic characteristics of Aerosil fumed silica (4th ed.), Tech. Bull. Fine Particles 11, Evonik Industries, Hanau, 2014.

Aerosil

-

Fumed

Silica.

Technical

Overview,

Evonik

Industries,

Hanau,

2015.

http://www.aerosil.com/product/aerosil/en/products/hydrophobic-fumed-silica/Pages/default.aspx. [12]

Cabot Corp., http://www.cabotcorp.com/solutions/products-plus/fumed-metal-oxides/hydrophobic

[13]

Wacker Chemie Ag., https://www.wacker.com/cms/en/products/brands_2/hdk/hdk.jsp

[14]

Hongwu International Group Ltd., http://www.hwnanomaterial.com/The-Difference-Between-HydrophilicSilica-and-Hydrophobic-Silica-Nanoparticles_p255.html

[15]

L. Wu, J. Baghdachi (Eds.), Functional Polymer Coatings: Principles, Methods, and Applications, First Edition, John Wiley & Sons, New York, 2015.

[16]

E.S.W. Kong (Ed.), Nanomaterials, Polymers and Devices: Materials Functionalization and Device Fabrication, Wiley, Hoboken NJ, 2015.

[17]

M. Okamoto, K. Miyazaki., A. Kado, S. Suzuki, E. Suzuki, Deoligomerization of cyclooligosiloxanes with dimethyl carbonate over solid-base catalysts, Catal. Lett. 88 (2003) 115-118.

[18]

T. Sabu, S. Ranimol (Eds.), Rubber Nanocomposites: Preparation, Properties, and Applications, John Wiley & Sons, Singapore, 2010.

[19]

X. Zhang, Y. Guan, Y. Zhao, Z. Zhang, D. Qiu, Reinforcement of silicone rubber with raspberry-like SiO2@polymer composite particles, Polym. Int. 64 (2015) 992–998.

[20]

L. Wang, X. Han, J. Li, X. Zhan, J. Chen, Hydrophobic nano-silica/polydimethylsiloxane membrane for dimethylcarbonate-methanol separation via pervaporation, Chem. Eng. J. 171 (2011) 1035–1044.

[21]

V.P. Silva, M.C. Gonçalves, I.V.P. Yoshida, Biogenic silica short fibers as alternative reinforcing fillers of silicone rubbers, J. Appl. Polym. Sci. 101 (2006) 290–299.

[22]

Q. Liu, J. Ding, D.E. Chambers, S. Debnath, S.L. Wunder, G.R. Baran, Filler-coupling agent-matrix interactions in silica/polymethylmethacrylate composites, J. Biomed. Mater. Res. 57 (2001) 384–393.

[23]

S. Thomas, R. Stephen (Eds.), Rubber Nanocomposites: Preparation, Properties, and Applications, John Wiley & Sons, Singapore, 2010. DOI: 10.1002/9780470823477.

18 [24]

M. Zupančič, M. Steinbücher, P. Gregorčič, I. Golobič, Enhanced pool-boiling heat transfer on laser-made hydrophobic/superhydrophilic polydimethylsiloxane-silica patterned surfaces, Appl. Therm. Eng. 91 (2015) 288–297.

[25]

M. Pantoja, J. Abenojar, M.A. Martinez, Influence of the type of solvent on the development of superhydrophobicity from silane-based solution containing nanoparticles, Applied Surface Science 397 (2017) 87–94.

[26]

C.K. Söz, E. Yilgör, I. Yilgör, Influence of the average surface roughness on the formation of superhydrophobic polymer surfaces through spin-coating with hydrophobic fumed silica, Polym. (United Kingdom) 62 (2015) 118–128.

[27]

C.K. Söz, E. Yilgör, I. Yilgör, Influence of the coating method on the formation of superhydrophobic siliconeurea surfaces modified with fumed silica nanoparticles, Prog. Org. Coatings. 84 (2015) 143–152.

[28]

F. Dolatzadeh, M.M. Jalili, S. Moradian, Influence of various loadings of hydrophilic or hydrophobic silica nanoparticles on water uptake and porosity of a polyurethane coating, Mater. Corros. 64 (2013) 609–618.

[29]

V.M. Gun'ko, V.V. Turov, V.M. Bogatyrev, B. Charmas, J. Skubiszewska-Zięba, R. Leboda, S.V. Pakhovchishin, V.I. Zarko, L.V. Petrus, O.V. Stebelska, M.D. Tsapko, Influence of partial hydrophobization of fumed silica by hexamethyldisilazane on interaction with water, Langmuir 19 (2003) 10816-10828.

[30]

A. Kawamura, C. Takai, M. Fuji, T. Shirai, Effect of solvent polarity and adsorbed water on reaction between hexyltriethoxysilane and fumed silica, Colloids Surfaces A: Physicochem. Eng. Asp. 492 (2016) 249–254.

[31]

B. Hamdi, T. Gottschalk-Gaudig, H. Balard, E. Brendlé, N. Nedjari, J.B. Donnet, Ageing process of some pyrogenic silica samples exposed to controlled relative humidities. Part I: Kinetic of water sorption and evolution of the surface silanol density, Colloids Surfaces A: Physicochem. Eng. Asp. 491 (2016) 62–69.

[32]

G. Lefebvre, L. Galet, A. Chamayou, Dry coating of talc particles with fumed silica: Influence of the silica concentration on the wettability and dispersibility of the composite particles, Powder Technol. 208 (2011) 372– 377.

[33]

H.N. Pham, A.E. Anderson, R.L. Johnson, K. Schmidt-Rohr, A.K. Datye, Improved hydrothermal stability of mesoporous oxides for reactions in the aqueous phase, Angew. Chemie - Int. Ed. 51 (2012) 13163–13167.

[34]

R. Marunaka, M. Kawaguchi, Rheological behavior of hydrophobic fumed silica suspensions in different alkanes, Colloids Surfaces A: Physicochem. Eng. Asp. 456 (2014) 75–82.

[35]

Y.-I. Kataoka, M. Kawaguchi, Effects of solvent power on dynamic moduli of polydimethylsiloxane-treated fumed silica suspension gels, Colloids Surfaces A: Physicochem. Eng. Asp. 436 (2013) 1041–1047.

[36]

J.N. Lee, C. Park, G.M. Whitesides, Solvent compatibility of poly(dimethylsiloxane)-based microfluidic devices, Anal. Chem. 75 (2003) 6544–6554.

[37]

J. Chen, Z. Ling, X. Fang, Z. Zhang, Experimental and numerical investigation of form-stable dodecane/hydrophobic fumed silica composite phase change materials for cold energy storage, Energy Convers. Manag. 105 (2015) 817–825.

[38]

C. Aydoğan, Z. El Rassi, Monolithic stationary phases with incorporated fumed silica nanoparticles. Part II. Polymethacrylate-based monolithic column with “covalently” incorporated modified octadecyl fumed silica nanoparticles for reversed-phase chromatography, J. Chromatogr. A 1445 (2016) 62-67.

[39]

A. Lazar, S.C. George, P.R. Jithesh, C.P. Vinod, A.P. Singh, Correlating the role of hydrophilic/hydrophobic nature of Rh(I) and Ru(II) supported organosilica/silica catalysts in organotransformation reactions, Appl. Catal. A Gen. 513 (2016) 138–146.

19 [40]

J.C. Cruz, P.H. Pfromm, J.M. Tomich, M.E. Rezac, Conformational changes and catalytic competency of hydrolases adsorbing on fumed silica nanoparticles: II. Secondary structure, Colloids Surfaces B Biointerfaces 81 (2010) 1–10.

[41]

J.C. Cruz, P.H. Pfromm, J.M. Tomich, M.E. Rezac, Conformational changes and catalytic competency of hydrolases adsorbing on fumed silica nanoparticles: I. Tertiary structure, Colloids and Surfaces B: Biointerfaces 79 (2010) 97–104.

[42]

V.M. Gun’ko, V.V. Turov, Nuclear Magnetic Resonance Studies of Interfacial Phenomena, CRC Press, Boca Raton, 2013.

[43]

V.M. Gun’ko, V.V. Turov, V.I. Zarko, O.V. Goncharuk, E.M. Pahklov, J. Skubiszewska-Zięba, J.P. Blitz, Interfacial phenomena at a surface of individual and complex fumed nanooxides, Adv. Colloid Interface Sci. 235 (2016) 108–189.

[44]

V.M. Gun’ko, V.I. Zarko, O.V. Goncharuk, A.K. Matkovsky, O.S. Remez, J. Skubiszewska-Zięba, G. Wojcik, B. Walusiak, J.P. Blitz, Nature and morphology of fumed oxides and features of interfacial phenomena, Appl. Surf. Sci. 366 (2016) 410–423.

[45]

V.M. Gun’ko, O.V. Goncharuk, J. Goworek, Evaporation of polar and nonpolar liquids from silica gels and fumed silica, Colloids Surf. A: Physicochem. Eng. Aspects 474 (2015) 52–62.

[46]

M. Parvinzadeh, S. Moradian, A. Rashidi, M.-E. Yazdanshenas, Surface characterization of polyethylene terephthalate / silica nanocomposites, Appl. Surf. Sci. 256 (2010) 2792–2802.

[47]

T. Bharathidasan, S.V. Kumar, M.S. Bobji, R.P.S. Chakradhar, B.J. Basu, Effect of wettability and surface roughness on ice-adhesion strength of hydrophilic, hydrophobic and superhydrophobic surfaces, Appl. Surf. Sci. 314 (2014) 241–250.

[48]

A.B. Beltran, G.M. Nisola, E. Cho, E.E.D. Lee, W.-J. Chung, Organosilane modified silica / polydimethylsiloxane mixed matrix membranes for enhanced propylene / nitrogen separation, Appl. Surf. Sci. 258 (2011) 337–345.

[49]

A.W. Adamson, A.P. Gast, Physical Chemistry of Surface, Sixth edition, Wiley, New York, 1997.

[50]

S.J. Gregg, K.S.W. Sing, Adsorption, Surface Area and Porosity, Academic Press, London, 1982.

[51]

V.M. Gun'ko, Composite materials: textural characteristics, Appl. Surf. Sci. 307 (2014) 444–454.

[52]

A.V. Kiselev, V.I. Lygin, Infrared Spectra of Surface Compounds, Wiley, New York, 1975.

[53]

M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery, Jr., J.E. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, O. Farkas, J.B. Foresman, J.V. Ortiz, J. Cioslowski, D.J. Fox, Gaussian 09, Revision D.01, Gaussian, Inc., Wallingford CT, 2013.

[54]

J.-D. Chai, M. Head-Gordon, Long-range corrected hybrid density functionals with damped atom-atom dispersion corrections, Phys. Chem. Chem. Phys. 10 (2008) 6615-6620.

20 [55]

K. Yang, J. Zheng, Y. Zhao, D.G. Truhlar, Tests of the RPBE, revPBE, τ-HCTHhyb, ωB97X-D, and MOHLYP density functional approximations and 29 others against representative databases for diverse bond energies and barrier heights in catalysis, J. Chem. Phys. 132 (2010) 164117.

[56]

A.V. Marenich, C.J. Cramer, D.G. Truhlar, Universal solvation model based on solute electron density and on a continuum model of the solvent defined by the bulk dielectric constant and atomic surface tensions, J. Phys. Chem. B 113 (2009) 6378-6396.

[57]

A.A. Granovsky, Firefly version 8.2.0, www http://classic.chem.msu.su/gran/firefly/index.html, Sep. 19, 2016.

[58]

A.D. Becke, Perspective: Fifty years of density-functional theory in chemical physics, J. Chem. Phys. 140 (2014) 18A301.

[59]

B. McCool, L. Murphy, C.P. Tripp, A simple FTIR technique for estimating the surface area of silica powders and films, J. Colloid Interface Sci. 295 (2006) 294–298.

[60]

V.M. Gun’ko, E.M. Pakhlov, J. Skubiszewska-Zięba, J.P. Blitz, Infrared spectroscopy as a tool for textural and structural characterization of individual and complex fumed oxides, Vibrational Spectroscopy 88 (2017) 56– 62.

[61]

C.L. Yaws (Ed.), Thermophysical Properties of Chemicals and Hydrocarbons Norwich, William Andrew Inc., New York, 2008.

21

Fig. 1. Incremental pore size distributions (SCV/SCR method) for (a) initial and dimethyl silylated silicas; (b) initial and trimethyl silylated silicas; (c) initial silica A-300, completely hydrophobized A-200 (AM1) and their mechanical blends as 3 : 1, 1 : 1, and 1 : 3; and (d) initial silica/titania ST20 (20 wt.% TiO 2) and ST76 (76 wt.% TiO2) and modified by HMDS or PDMS/DMC. Inserts are HRTEM images (scale bar 20 nm) of initial silica A-300 and ST20.

22

Fig. 2. Particle size distributions of silicas in aqueous (dashed lines) and aqueous-ethanol (solid lines) media with respect to particle (a, c, e) volume and (b, d, f) number for: (a, b) first, (c, d) second, and (e, f) third series of sample s. Inserts show SEM images of A-300 corresponding to aggregates (a) and agglomerates (b).

23

Fig. 3. Evaporation kinetics of ethanol (0.5 g per 0.05 g of oxide) from suspensions containing silicas as a relative amount mev/m0 (where mev and m0 are the evaporated and initial ethanol masses, respectively, recalculated per 0.05 g of silica) divided by SBET for series: (a) M0-M5; (b) CM1-CM4 and A-300; and (d) B1-B3, A-300, and AM1 over the 1350-1900 min range (full curves are shown in Fig. S11) corresponding to evaporation of weakly (WBE) and strongly (SBE) bound ethanol.

24

Fig. 4. Solvation energy vs. a number of TMS groups in clusters Si8O12(OH)n(OSi(CH3)3)8-n (n = 8, 7, 6,… 0) calculated using SMD/B97X-D/cc-pVDZ//cc-pVDZ (see Table S2 and Fig. S12 in SI file) for water, ethanol, and n-hexane as solvents.

25

Fig. 5. DSC curves for water interacting with silicas of the (a, b) first, (c, d) second, and (e, f) third series of samples upon (a, c, e) cooling and (b, d, f) heating. (b) Curves for decane (dot-dashed lines) are shown at the scale (T-Tm).

26

Fig. 6. DSC curves for n-decane interacting with silicas of the (a, b) first, (c, d) second, and (e, f) third series of samples upon (a, c, e) cooling and (b, d, f) heating.

27

Fig. 7. DSC curves for (a, c) water and (b, d) n-decane interacting with (a, b) ST20 and (c, d) ST76 upon cooling and heating. (d) Curves for decane interacting with AST71 are shown.

28

Fig. 8. Size distributions of pores filled by liquid (a, b, e) water or (c, d, f) n-decane according to thermoporometry calculations with DSC melting curves.

29 Table 1 Textural characteristics of initial and differently modified silicas (SCV/SCR method). Sample

MS (%)

M0

0

SBET 2

Snano 2

Smeso 2

Smacro 2

Vp 3

Vnano

Vmeso

3

3

Vmacro 3



(m /

(m /

(m /

(m /

(cm /

(cm /

(cm /

(cm /

>

>

g)

g)

g)

g)

g)

g)

g)

g)

(nm)

(nm)

234

35

191

8

0.602

0.019

0.468

0.116

16.8

5.42

3

w

0.06 6

M1

22.5

235

30

187

18

0.756

0.016

0.438

0.301

24.1

6.99

9

0.24 5

M2

50.0

231

26

197

9

0.630

0.014

0.475

0.141

18.2

5.80

2

0.07 7

M3

68.7

225

19

200

6

0.593

0.011

0.484

0.098

M4

89.4

217

14

199

5

0.573

0.008

0.494

0.071

15.8

5.61

8 14.2

9 5.49

7 M5

100.

216

11

201

4

0.555

0.007

0.485

0.063

262

49

205

8

0.621

0.025

0.470

0.126

0 A-300

0

13.4

0.08 6

5.42

8 17.1

0.00

0.12 4

5.02

4

0.13 6

CM1

20.0

244

36

200

8

0.613

0.019

0.476

0.118

17.0

5.33

5

0.09 4

CM2

35.0

236

30

198

8

0.621

0.016

0.476

0.128

17.1

5.56

6

0.08 4

CM3

73.0

224

20

194

10

0.650

0.011

0.487

0.152

18.4

6.19

5

0.06 2

CM4

100.

212

12

191

9

0.642

0.007

0.502

0.133

0

17.2

6.38

7

0.00 5

B1

25.0

238

36

196

6

0.563

0.019

0.457

0.087

15.3

4.97

2

0.05 9

B2

50.0

220

25

187

9

0.626

0.014

0.467

0.146

18.4 5

6.06

0.08 7

30 B3

75.0

198

15

175

8

0.565

0.009

0.433

0.123

17.4

6.10

4

0.06 8

AM1

100.

178

8

166

5

0.474

0.005

0.399

0.070

0 ST20

-

14.9

5.78

8 84

9

70

3

0.179

0.008

0.147

0.062

22.4

0.06 9

5.46

6

0.34 4

ST20/HMDS

100

69

0

62

7

0.269

0.0

0.163

0.106

26.7

10.2

-

6

9

0.16 8

ST20/PDMS/D

100

36

0

17

19

0.182

0.0

0.038

0.145

MC ST76

-

22

4

17

1

0.056

0.002

0.032

0.022

40.0

15.3

0.17

6

7

1

24.4

5.53

3

0.41 0

ST76/HMDS

100

20

0

17

3

0.084

0.0

0.027

0.057

44.0

11.4

-

9

5

0.40 3

ST76/PDMS/D

100

8

0

7

1

0.023

0.0

0.008

0.014

MC

46.5

10.2

-

2

5

0.59 1

Note. Series M0-M5 and AM1 are commercial samples (PPCISC); series CM1-CM4 (lab samples with A-300 modified by hexamethyldisilazane), and samples B1, B2, and B3 are mechanical blends of A-300 with AM1 at the weight ratio 3 : 1, 1 : 1, and 1 : 3, respectively. MS is the degree of substitution of surface silanols by hydrophobic groups. The values of Vnano and Snano were calculated by integration of the fV(R) and fS(R) functions, respectively, at 0.35 nm < R < 1 nm, Vmeso and Smeso at 1 nm < R < 25 nm, and Vmacro and Smacro at 25 nm < R < 100 nm.